|analytical Q||Suggest a Link||Contact||Search||Energy|
Final published article:
PDF (558 kB)
html on Platts.com
Recommend this page to a friend:
Risk management and trading ASP's to the rescue
by Anne Ku (November 2001)
original draft of article published in November/December 2001 issue of Global Energy Business
Technology Or Risk Management
The need for speed: getting risk management asap via ASP Trading and
Systems for trade-capture, position keeping, and risk management have always resided in-house not just because of their mission critical nature. Until the Internet proved its reliability as a distribution channel for software, trading companies who choose not to build in-house had to go through the painful and lengthy process of system selection and implementation. Over time, this process became an accepted evil: sinking your money into a deep black hole to buy a big black box.
Application service providers (ASP) now offer the functionality of such systems via the Internet on a pay as you go basis. ASPs attract several types of customers: those that are still relying on spreadsheets to capture their deals and track their risk; those with low energy transactions such as major energy buyers of industrial companies or certain utilities, those who are unhappy (or need to replace) with in-house developed legacy systems cobbled over time that are no longer adequate, and the curious but dissatisfied user who wants to "try before you buy."
Not a new invention
Since the Internet made its debut in the business world, the technology to deliver applications over the Internet has spun into so-called Application Service Providers (ASP). Last year, Energy IT, a sister magazine, presented solid reasons why utilities should use ASP's and why there was still resistance. "The ASP model represents a paradigm shift in the concept of software distribution - from that of a product to that of a service." It lets the client "rent" the application and provides a "quick-fix" to solutions that would otherwise take a long time to build.
In the financial sector, ASP's promises to free companies from the hassle and expense of maintaining trading and risk management systems which are built on client-server architecture. DerivativesStrategy.com in February 2000 reports "derivatives software vendors are frantically transforming themselves into ASPs in order to offer their high-priced client-server packages to a wider user base."
The need for speed
In these days of cell phones and instant messaging, if you don't respond like you are completely wired, then you'd miss the opportunity. Companies are willing to reorganize to respond to market changes. Yet, people are willing to wait months to years for a mission critical system to get implemented. And users wonder why in-house help desks spend more time on "loging" the request than fixing their problems right away.
If there's an alternative available it would be a system that is ready, with full functionality, instant upgrades, 24 by 7 expert support and responsiveness to users, secure, scaleable, and intuitive to use (minimal training time). But does such a solution exist?
A sense of urgency
There's an urgency to get risk management under control, says Phil Inje Chang, president of San Francisco-based e-acumen. Top tier players can afford to build in-house or purchase large systems that get continuously enhanced. With Internet technology, others can now go for ASPs which require no upfront fee or lengthy implementation period. He observes that people are still sensitive to moving portfolio data outside of internal network. As such, their acuRisk product is not available as an ASP (but the analytics used for market intelligence: acupower, kiloweather, and powerview are.) He thinks it will get there eventually - as this is more about perception of security than the development effort.
A level playing field
That big players can afford highly sophisticated systems while the smaller ones have to rely on home-spun, piecemeal spreadsheets was a truth many accepted. Until the Internet arrived. Or more specifically, until ASP's arrived.
As a former Goldman Sachs trader, Raj Mahajan, co-founder and president of New York-based Kiodex, noticed that his clients were often ill-equipped with information compared to their larger counterparties. This asymmetric access to information translated to a lack of neutrality and a non-level playing field. Large energy-users such as chemical companies, pulp and paper, food, tobacco, and airlines are primarily hedgers and as such, do not have the large sophisticated trading/risk management systems of large energy trading houses. With the Internet as a distribution channel, an ASP can provide immediate and affordable access to the same kinds of tools. Kiodex's ASP called Risk WorkBench, which provides deal capture, pricing, risk reports, etc., is undergoing a beta test by a cross section of corporate energy end-users and utilities, some of which may be live by the time you read this.
A trader's dream come true
The founders of New York-based Sakonnet Technology were frustrated and dissatisfied users at JP Morgan. Thurstan Bannister, president, who traded metals, crude, financial gas, and exotic options, recalls missing a good trade because it could not be captured by the system. He says, "systems have been a source of restraint on going to markets. They are an annoyance - a necessary evil at best and a nightmare at worst."
Traders resent having to stay back late because end-of-processes take too long to get meaningful position reports. They want to run analysis in real-time. There is too much uncertainty in waiting for upgrades to in-house systems. Given the fast trading environment, you can't afford to wait. Bannister adds, the biggest problem today is that existing systems cannot keep up with business requirements.
He wanted a system which is low cost, has instant expert support (not a helpdesk who needs to refer to someone else), and high adaptability to products and markets. He and his colleagues (the technologists) who have worked with third-party and in-house systems got together to create just that. Xenon, now in its third release, in Java and XML, is delivered to clients over the Internet or housed as an Intranet application.
Sakonnet's first client was London-based energy derivatives trader Francis Cicoli-Abad who recalls specifying his requirements for gas options and getting the beta release within the first 1.5 months. Within four months, he got the alpha version and became a paying customer. Although his management had already committed to a third-party client-server system which would not be customised or implemented for a while, he convinced them to let him use Xenon as a stopgap measure. A year and a half and some thirty software modifications/upgrades later, he's moved onto another energy company. But the team he left behind is still using Xenon. And at his new company, he is Sakonnet's first return customer.
Cicoli-Abad says that it's not the lock-in of the system but rather the (unparalled) unsurpassable response time of the ASP. Sakonnet's experts consistently beat in-house support in his call for help. Traders can't afford to lose more than a couple hours in the day sitting on positions that include options which are very dynamic. He must be able to review it constantly with scenarios. Excel is simply not strong enough to handle real-time scenarios where you need to actively manage your position over the life of the product.
Sunk cost and stop loss (OR the big black hole for the big black box OR "..know when to fold 'em ")
So it is easy to move to an ASP if you are currently using spreadsheets. But if you're using an expensive system that you've invested a lot of time, money, and effort in specifying your requirements, selecting a system, building/customizing, integrating, and implementing it? You can hardly walk away and grab an ASP with more current functionality.
In economics, it's called sunk cost. In accounting terms, it's a write-off. A good trader, says Cicoli-Abad, is one who knows when to quit. In trading terms, it's called stop loss.
But the joint decisions of front office, mid office, IT, and other departments and the politics make such efforts (see box) difficult to walk away from. Ashley Abbott, Financial Engineering VP at ForwardVue Technologies, Austin, Texas, observes that most people end up buying a big black box. That's the problem. In her previous position as managing director of global trading and risk management at a top tier energy company, she witnessed the organizational behavior in decisions on such a highly visible project. "You end up choosing a system that is least bad. You also run into a conflict of interest." She says, there are plenty of third-party vendors, but few have what large companies need. There is no such thing as a pure build or buy. The result is always a hybrid and you have to commit to doing both. People get hung up on details of requirements and documentation - and this has a spiralling effect.
Rumours abound of systems that take ages to implement because of requirements that keep changing. When you finally get the long awaited system, how can you not use it?
Incumbents slow to react
Given the lengthy time to customisation, integration and implementation of highly sophisticated systems, some customers are still prepared to wait. These are the very ones who are skeptical of ASP's ability to deliver equally sophisticated systems equally fast.
When asked whether he sees the ASP model for his company, Coleman Fung, President, New York-based OpenLink, replies "OpenLink has the technology and the tools to provide ASP-based services, but doesn't see any 'real' market demand for such services. If market demand reaches critical mass, OpenLink can introduce these services fairly quickly. However, many trading-oriented energy firms view managing their internal trading and risk management systems as strategic and critical. The best audience for ASP-based services is probably energy consumers such as large industrial companies and smaller utilities, but we just don't see any significant demand from these entities. Perhaps many low-end products are currently addressing the needs of these companies."
Another non-ASP risk management systems provider KWI plans to introduce new product functionality in 1Q 2002 that will include web-based components to operate in concert with their main product kW3000(TM). Steve Delaney, North American Sales Director, KWI, Houston, says "these web-based products will use the Stochastic Price Shape (SPS) model for hourly price shape development and OASIS monitoring for transmission congestion identification, among other tools. This will enhance the front-to-back end processing capability of the kW3000 (TM) in handling physical power delivery."
Blake Pound, Partner at Accenture's Energy Practice in Houston, explains the reluctance of incumbents to rush into converting their existing products to ASP. Their revenue models are built around selling licenses, maintenance, support, and consulting. It would take a mindset and strategic change for these vendors to move away from this to a subscription-based or transaction volume-based revenue model. They could develop ASP versions of their software if the clients demands yet. The upper tiered energy trading companies, with their growing trading volumes, require scaleability (see box) - which is something not yet proven with ASP. Trading organizations have a higher sense of uneasiness about the possibility of trading data getting out - as such, people have a hard time getting used to the notion of having their data reside on a server that is physically outside their building. [But September 11th could change that.] Bigger trading companies also have concerns about integration with other systems. How do you integrate with in-house systems if it's hosted elsewhere?
To allow its software to reach a wider audience, TradeCapture has converted its client-server architecture software to a web browser-based ASP. ICTS-online is charged as a pay as you go based on transaction volume with fixed fees for data changes. Matt Frye, Chief Marketing Officer, summarizes: if you're big, you'd want a client-server based software or an ASP-product hosted in-house on an intranet. If you're small, you would choose an ASP. There's a diminishing marginal return for software vendors to commoditise their product for all if there are only ten multi-commodity trading companies that want it. However, there are plenty of mid to smaller sized companies who didn't think they could afford sophisticated tools before ASPs appeared on the scene.
Developed in-house, spun out as ASP
While existing third-party vendors may not see the value of ASP yet, two energy companies have taken their in-house systems to their clients as ASP. Enron started marketing its Enron Energy Desk as early as 1999 when it was formed in Europe. Primarily active in the Nordic area, EnergyDesk.Com was conceived originally to stimulate deregulation and liquidity and facilitate trading of non-standard products. The rationale was that it would result in higher liquidity, lower transaction costs for all players, and a more competitive market. The service is at present actively marketed in the Nordic area and for power only. Other geographical markets which display similar characteristics to the Nordic market will be considered in the future. Their target clients include trading companies, generators, utilites and industrials - with a focus on companies that actively manage their risk exposure. Before taking on their ASP product RiskDesk, many were using spreadsheets and competitor systems delivered in the traditional way, not ASP. Resistance to ASP is almost non-existent in the Nordic area. The benefits of reduced costs, 24 hour support, reliability and freqent release cycles seem to be embraced by the large majority of players in the area. The revenue model is based on monthly subscription, where the fee is based on the length of the contract and number of users.
In late August, Aquila announced its version of a risk management ASP. Costing from $6,000 a month, Risk180 (see figure) is being marketed to utilities, municipalities, energy aggregators and small energy marketers, as well as large industrial firms. These clients typically are low-volume traders that use spreadsheets to manage their energy transactions. The product was designed to meet the deal capture and portfolio valuation needs of clients who do not wish to purchase or build an expensive, enterprise-wide risk management software system. Developed jointly with with SunGard Trading and Risk Systems, Risk180 currently handles only gas transactions with daily reporting. Power will be added later - as is the aim for real time reporting.
Vice President Jennifer Fisher sees the real value of their ASP as two fold: the human service behind the software and access to curves of top five marketers. Clients don't know how to decompose risks or interpret risk exposures. They have over 100 mid-offie experts who can help them do that. While they don't release their forward curves, they do mark the clients' trades against them, giving a more accurate marked-to-market position than if they were to do so otherwise. The Radius Group is separated from Aquila by a Chinese Wall, that is, the trading floor does not have access to Risk180's client data.
There are five components of a risk management structure, she says: policy and process, organization structure, management reporting, model analysis, systems and data. Although Risk180 helps with the last three, you need all to work. She adds that such systems are mission critical, without which you could be exposed to negative risk (negative impacts of risk) or hidden risk (not capturing opportunity to make more money).
Try before you buy
One attraction of the ASP model is that you can try before you buy as it does not require an upfront fee. This convenience factor is not available for systems in which you have to install in-house.
Sakonnet offers a two-week free trial period and claims that trial users almost always end up using them. Since this summer, they already have five clients who have agreed to "rent" their ASP.
Business intrusion and version control
While it is harder, if not impossible, to "try" a client-server based system, Ashley Abbott gives ways to test if a system does what you expect it to do. She would ask the vendors to test a portfolio with complex physical structures and see how fast they are able to figure it out. She would check how intrusive the system is to the business. Anyone can do financial products, it's the physical side that's difficult. For example, how do you eliminate Value Added Tax on gas transactions across Europe. How do you translate a pipeline structure from market to market?
An ASP is less intrusive in that upgrades can be done without the need to send someone physically to go around each desk and install the upgrades. Version control can be done remotely and seamlessly. But this does not preclude the integration problem, says Blake Pound. Typically data model changes (which could be due to technology upgrades) could also break in-house integration with other systems, if there is no forewarning. As clients generally want to be on the latest release, it requires effort to give advance warning in order to continually upgrade and make modifications to other systems.
Upgrades are necessary as the energy world is changing so quickly. Traded products become more complicated and varied while model assumptions may no longer hold. A former interest rate options trader Michael Sigman, Executive Director, Capstone Global Energy LLC, Houston, Texas, says that he's never come across a modelling environment as challenging as one for energy in general, with electricity in particular. Because ASP software is built on the newer Internet technology, it doesn't have million lines of code that touch different modules - hence, changes can be done more easily. But he's yet to see software that can incorporate options and other elements of a portfolio such as customer contracts and real assets (including how to value storage and transmission contracts) into a single unified consolidated risk framework. Before you even decide on a system, you need to define your porfolio and modelling framework, he advises.
Stop gap or long term solution
A recent E-source report (see box) on ASPs for energy service providers urges companies to determine a realistic return on investment beforehand. "Just as leasing an automobile can ultimately be more expensive than buying - the initial cost is usually much lower than any other alternative, but an ASP may cost more in the long run." (next two sentences are mine) However, if the markets keep changing and business requirements follow suit, an ASP which offers frequent upgrades may not be a bad idea at all. After all, who wants to drive an old car if there are new releases every month? Among other things, it advises to check the provider's financial strength, as "any good, well-run ASP should be profitable, because the ASP model is a high-margin business."
Making it work
To make the ASP solution work, the report suggests two things. One, determine what problem needs to be fixed before approaching the ASP. [next are my words] For those companies currently dependent on in-house spreadsheets to capture their deals, track their positions, and produce risk reports, moving to an ASP seems an obvious solution. For those companies trapped in legacy systems (whether in-house or third-party installations), the migration to an ASP may be a trickier decision because of the psychology of "sunk cost." Some companies (like Statoil) have chosen (choose) an ASP as a stop gap measure, until the new system gets integrated and implemented. "A big question for those looking for a risk management system is defining the boundaries of the system," says Don Winslow, VP Mid Office, Pacificorps, Portland, Oregon. "Does it include scheduling and back office? No vendor can deliver everthing. It is my experience that the greatest complications arise for scheduling and back office, not in risk management. The same principles work in risk management whether you are in the US or Europe. However, each jurisdiction has unique requirements for back office and scheduling."
Two, build a partnership with your ASP - outside the traditional client/vendor type relationship. [next are my words] Consider the trader at Statoil who worked with Sakonnet Technology to develop the functionalities for managing his natural gas options risks, he is now advocating this ASP in the new firm he's joined. Consider Toronto-based Algorithmics which has several ASP partnerships with financial institutions and exchanges.
Table: pros and cons of using ASP
URLs of companies mentioned in this article